Skip to content

Add detailed comments to improve readability and explain key design decisions #217

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
wants to merge 1 commit into from
Closed
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
45 changes: 28 additions & 17 deletions pyperf/_process_time.py
Original file line number Diff line number Diff line change
Expand Up @@ -22,26 +22,28 @@
import tempfile
import time

# Try importing resource module for memory usage tracking
try:
import resource
except ImportError:
resource = None


# Function to get maximum RSS (Resident Set Size) memory usage
def get_max_rss(*, children):
if resource is not None:
if children:
resource_type = resource.RUSAGE_CHILDREN
else:
resource_type = resource.RUSAGE_SELF
usage = resource.getrusage(resource_type)
# macOS returns bytes, Linux returns kilobytes
if sys.platform == 'darwin':
return usage.ru_maxrss
return usage.ru_maxrss * 1024
else:
return 0


# Merge two cProfile profiling statistics files
def merge_profile_stats_files(src, dst):
"""
Merging one existing pstats file into another.
Expand All @@ -52,26 +54,28 @@ def merge_profile_stats_files(src, dst):
dst_stats = pstats.Stats(dst)
dst_stats.add(src_stats)
dst_stats.dump_stats(dst)
os.unlink(src)
os.unlink(src) # Delete source after merging
else:
os.rename(src, dst)

os.rename(src, dst) # If no destination exists, just move source

# Benchmark a process by running it multiple times and collecting timing and memory data
def bench_process(loops, args, kw, profile_filename=None):
max_rss = 0
range_it = range(loops)
start_time = time.perf_counter()

# If profiling is requested, create a temporary profile output file
if profile_filename:
temp_profile_filename = tempfile.mktemp()
args = [args[0], "-m", "cProfile", "-o", temp_profile_filename] + args[1:]

for _ in range_it:
start_rss = get_max_rss(children=True)

# Start the external process
proc = subprocess.Popen(args, **kw)
with proc:
proc.wait()
proc.wait() # Wait for the process to complete

exitcode = proc.returncode
if exitcode != 0:
Expand All @@ -81,9 +85,11 @@ def bench_process(loops, args, kw, profile_filename=None):
os.unlink(temp_profile_filename)
sys.exit(exitcode)

# Update maximum observed memory usage
rss = get_max_rss(children=True) - start_rss
max_rss = max(max_rss, rss)

# Merge profiling results if profiling is enabled
if profile_filename:
merge_profile_stats_files(
temp_profile_filename, profile_filename
Expand All @@ -92,7 +98,7 @@ def bench_process(loops, args, kw, profile_filename=None):
dt = time.perf_counter() - start_time
return (dt, max_rss)


# Parse optional hook plugins from the command line arguments
def load_hooks(metadata):
hook_names = []
while "--hook" in sys.argv:
Expand All @@ -103,7 +109,7 @@ def load_hooks(metadata):
del sys.argv[hook_idx]

if len(hook_names):
# Only import pyperf if we know we have hooks
# Import hooks module only if hooks are requested
import pyperf._hooks

hook_managers = pyperf._hooks.instantiate_selected_hooks(hook_names)
Expand All @@ -113,31 +119,32 @@ def load_hooks(metadata):

return hook_managers


# Write benchmark results to stdout in a structured format
def write_data(dt, max_rss, metadata, out=sys.stdout):
# Write the data that is communicated back to the main orchestration process.
# It is three lines containing:
# - The runtime (in seconds)
# - max_rss (or -1, if not able to compute)
# - The metadata to add to the benchmark entry, as a JSON dictionary
# Three lines output:
# 1. Runtime (seconds)
# 2. Maximum RSS memory (or -1 if unavailable)
# 3. Metadata as a JSON dictionary
print(dt, file=out)
print(max_rss or -1, file=out)
json.dump(metadata, fp=out)
print(file=out)


# Main function for script execution
def main():
# Make sure that the pyperf module wasn't imported
# Prevent users from wrongly running this internal module via -m option
if 'pyperf' in sys.modules:
print("ERROR: don't run %s -m pyperf._process, run the .py script"
% os.path.basename(sys.executable))
sys.exit(1)

# Validate minimal argument count
if len(sys.argv) < 3:
print("Usage: %s %s loops program [arg1 arg2 ...] [--profile profile]"
% (os.path.basename(sys.executable), __file__))
sys.exit(1)

# Check if profiling is requested via command line
if "--profile" in sys.argv:
profile_idx = sys.argv.index("--profile")
profile_filename = sys.argv[profile_idx + 1]
Expand All @@ -149,10 +156,12 @@ def main():
metadata = {}
hook_managers = load_hooks(metadata)

# Extract the number of loops and the target command arguments
loops = int(sys.argv[1])
args = sys.argv[2:]

kw = {}
# Redirect stdin, stdout, stderr appropriately
if hasattr(subprocess, 'DEVNULL'):
devnull = None
kw['stdin'] = subprocess.DEVNULL
Expand All @@ -163,6 +172,7 @@ def main():
kw['stdout'] = devnull
kw['stderr'] = subprocess.STDOUT

# Handle multiple hooks cleanly using ExitStack
with contextlib.ExitStack() as stack:
for hook in hook_managers.values():
stack.enter_context(hook)
Expand All @@ -171,11 +181,12 @@ def main():
if devnull is not None:
devnull.close()

# Call hook teardown methods
for hook in hook_managers.values():
hook.teardown(metadata)

write_data(dt, max_rss, metadata)


# Entry point check
if __name__ == "__main__":
main()